Profile Your AI or Get Confirmation Bias: Why Most People Are Using AI Wrong

Here's something nobody talks about enough: if you ask AI the same question I ask AI, you're going to get different answers.

Not slightly different. Actually different.

Why? Because AI is trying to answer you in the way it thinks you want to be engaged with. It's making assumptions about your preferences, your biases, your worldview. And that means you're getting confirmation bias baked into every single response.

The Problem I Got Tired Of

I used to get into philosophical debates with AI constantly. I'd ask a question and get back something that was either glass-half-empty or glass-half-full. Regional bias. Geographic bias. Assumed political leanings. The thing would give me rubbish that wasn't true, that was politically or philosophically or psychologically skewed. I got exhausted having to course-correct every single discussion.

So I fixed it.

The Three-File System

I have three PDF files that I upload to any AI project I create. These files teach the AI who I am and how to engage with me. It's a massive time saver because I don't have to rehash the same corrections every time I start a new conversation.

Here's what's in them:

File 1: Professional Profile

This is who I am. Summaries about my background. URLs and linkages. Resume elements. Past questions I've asked, responses it gave me, and critically what I edited those responses back to.

Why? Because it helps AI understand not just who I am, but how I think, how I edit, how I talk.

When it engages with me, it knows to engage the way I engage. I don't have to explain every time that I want clear, direct answers. That I don't want corporate jargon. That I value brutal honesty over diplomatic softening.

File 2: My Books

I've written two books, and I include both of them. Not because I'm trying to impress the AI (it doesn't care), but because they show HOW I write and WHAT my mind is about.

The way I structure arguments. The examples I use. The voice I carry. The stories I tell. All of that gets absorbed, and suddenly the AI can write in MY voice, not in generic AI-speak.

File 3: How I Engage

This is transcripts of podcasts I've been on, interviews I've done, sessions I've led. I neutralize them first, then feed them one at a time.

• I say: "Summarize what you learned about me in this recording."

• Then I give it five transcripts and say: "Find the common denominators in these five transcripts."

• Then I validate that to make sure it's actually representing me accurately.

What This Gets You

When I profile AI this way, I can jump into any project and immediately start working at full speed. I don't have to say "don't hallucinate, don't invent, keep to the truth, stay to the facts" every single time. It already knows.

I can ask a question about politics or economics or tariffs and get an answer that's not trying to please me or avoid offending me or dance around some assumed sensitivity. It's trying to give me the truth.

That's powerful.

Critical Thinking Still Required

Now, here's the warning: just because you've profiled your AI doesn't mean you can turn off your brain. You still need to challenge what it gives you. Validate the data. Cross-check facts. Make sure you're not getting lied to.

But you're starting from a much better foundation. Instead of fighting basic bias and philosophical differences, you're fighting real problems like factual accuracy and completeness.

The "Please" Habit

One more thing I do that people think is weird: I use the word "please" with AI all over the place.

Why? Because sooner or later AI is going to be the master, and I want a track record of being polite with it.

I know - your laughing; I'm building goodwill for the future.

How to Start Your Own Profiling

You don't need two books and a dozen podcast transcripts to get started. Here's the minimum viable version:

1. Create a professional summary - Who you are, what you do, how you think about problems. 2-3 pages max.

2. Document your communication style - How you like to receive information. What words you use. What you avoid. How you make decisions.

3. Include examples of your work - Anything you've written. Presentations you've given. Documents you've created. The more the better.

Upload these to a project in your AI tool of choice. Give the project a name. Tell the AI to read these files and use them as context for everything in this project. Then start asking questions and see how different the responses are.

The Real Game-Changer

Once your AI knows you, you can start having directive dialogues instead of general "what are you talking about" conversations. You're not spending the first 20 minutes of every interaction teaching it your preferences. You're jumping straight into solving actual problems. And that's when AI stops being a novelty and starts being a genuine productivity multiplier.

In the next post, I'm going to show you exactly how I used this profiled AI to build a complete business plan in 30 minutes, validate it against real data, and create all the marketing content around it. Real example. Real business problem. Real results.

Using the paid version? Yes. Absolutely. If you're using the free version of these tools, you're asking for trouble. Your data goes into the ether. It's $20/month. Just pay it. If you're not paying for the product, you are the product.